I love AI; I hate AI (www.ewanvalentine.co.uk)

The author describes a love–hate relationship with developer AI tools like Cursor: they speed up tedious tasks (e.g., hunting typos in 500-line YAML) and free creators for higher‑level work, but simultaneously undermine the sense of software engineering as a craft and threaten careers. Citing real‑world signals — Amazon’s plan to cut roughly 14,000 corporate roles “to seize the opportunity provided by AI” — the piece argues that both knowledge and software engineers are at risk of being automated away, often by models that can replace creative, knowledge‑work faster than physical/manual tasks. The essay warns of concentrated technical and economic power — a few trillionaire owners controlling massive data‑centers and AI infrastructure — and the emergence of a techno‑feudalist dynamic where automation decisions are driven by billionaire incentives, not public interest. Technically, it highlights a counterintuitive trend: current AI is better at automating cognitive/creative tasks (LLMs, coding assistants) than many everyday bureaucratic or physical chores, leaving dull but essential services untouched. The implication for AI/ML is clear: we need policy, governance and new social models to rebalance who decides what gets automated, to mitigate labor displacement, monopoly risk, and to preserve meaningful work as automation accelerates.